Since Anfinsen's demonstration that the sequence of a protein dictates its structure, the energetics and kinetics of folding have been explored with ever-greater temporal and spatial resolution. However, almost all of these studies have been devoted to water-soluble proteins, and our understanding of membrane proteins is in its infancy. Here, we use protein design to test and refine our understanding of membrane protein folding. Model peptides are designed to associate into helical bundles in micelles and phospholipid bilayers. The study of these systems is simplified by the fact that the unfolded state is well defined (the monomeric helices), and the energetics of folding can be monitored by measuring the monomer-nmer equilibrium. In aims 1 &2, we test the importance and relative energetic contributions of disparate forces such as van der Waals packing versus polar interactions. These studies will not only further our understanding of protein folding, but also provide insight and methods to understand the process of transmembrane helix association, which often plays an essential role in signal transduction. As an outgrowth of our understanding of the folding of membrane proteins, we are beginning to design water-soluble versions of 2 membrane proteins, phospholamban and the KcsA potassium channel. These studies promise to provide a new method for obtaining large quantities of water-soluble versions of membrane proteins for pharmaceutical and biophysical studies. Our specific aims are as follows: Aim 1. We will study the association of designed and natural transmembrane peptides, and use these systems to understand the features leading to the folding and association of transmembrane helices. Aim 2. We will explore how regions of a protein located in the aqueous, headgroup, and membrane interior cooperate to dictate the fold of a protein by designing helical bundles that incorporate both transmembrane as well as water-soluble folding motifs. Aim 3 We will structurally characterize water-soluble versions of phospholamban and KcsA. Aim4. We will develop computational methods for transmembrane protein structure prediction &design.